Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

Dataset: 80% C4 (Web Text) and 20% Python Code

Neuron 2651 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1520. Min Act: -0.0000

Data Index: 1189933 (C4 (Web Text))

Max Activating Token Index: 156

Text #1

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1207. Min Act: -0.0000

Data Index: 676328 (C4 (Web Text))

Max Activating Token Index: 219

Text #2

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1118. Min Act: -0.0001

Data Index: 1187969 (C4 (Web Text))

Max Activating Token Index: 289

Text #3

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1144. Min Act: -0.0000

Data Index: 1602951 (Python Code)

Max Activating Token Index: 232

Text #4

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1177. Min Act: -0.0000

Data Index: 148432 (C4 (Web Text))

Max Activating Token Index: 156

Text #5

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1019. Min Act: -0.0000

Data Index: 627612 (C4 (Web Text))

Max Activating Token Index: 179

Text #6

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1016. Min Act: -0.0000

Data Index: 956237 (C4 (Web Text))

Max Activating Token Index: 110

Text #7

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1021. Min Act: -0.0001

Data Index: 409611 (C4 (Web Text))

Max Activating Token Index: 163

Text #8

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.0991. Min Act: -0.0001

Data Index: 858111 (C4 (Web Text))

Max Activating Token Index: 392

Text #9

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.0953. Min Act: -0.0000

Data Index: 630209 (C4 (Web Text))

Max Activating Token Index: 616

Text #10

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.0924. Min Act: -0.0000

Data Index: 282540 (C4 (Web Text))

Max Activating Token Index: 218

Text #11

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.0969. Min Act: -0.0000

Data Index: 1596397 (Python Code)

Max Activating Token Index: 567

Text #12

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.0949. Min Act: -0.0001

Data Index: 1549258 (Python Code)

Max Activating Token Index: 526

Text #13

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.0906. Min Act: -0.0000

Data Index: 1016600 (C4 (Web Text))

Max Activating Token Index: 585

Text #14

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1045. Min Act: -0.0000

Data Index: 580724 (C4 (Web Text))

Max Activating Token Index: 852

Text #15

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.0935. Min Act: -0.0001

Data Index: 887615 (C4 (Web Text))

Max Activating Token Index: 161

Text #16

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.0926. Min Act: -0.0000

Data Index: 811114 (C4 (Web Text))

Max Activating Token Index: 457

Text #17

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.0988. Min Act: -0.0000

Data Index: 1547535 (Python Code)

Max Activating Token Index: 427

Text #18

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.0996. Min Act: -0.0001

Data Index: 1366501 (Python Code)

Max Activating Token Index: 892

Text #19

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.0867. Min Act: -0.0000

Data Index: 1198879 (C4 (Web Text))

Max Activating Token Index: 573